promptdojo_

The harness-engineering mindset — Agent equals Model plus Harness — step 8 of 8

ch 30 · harness engineering8/8
Checkpoint

One last thing before we move on. Same surface as a write step — but the lesson doesn't complete until this passes.

Final drill for the mindset lesson. Write triage_harnesses(harnesses) that scores five harness configurations and returns a dict:

  • scores: dict mapping each harness name → its 0-100 score (insertion order preserved)
  • weakest: the name of the harness MOST likely to fail first (lowest score). On a tie, return the FIRST one in input order.

Reuse the six-piece scoring from the previous drill:

  • Each present piece (config, tools, infra, orchestration, hooks, observability) adds 16.
  • All six present → 100.

Five harnesses run for you. Expected output:

scores: {'claude-code-pro': 100, 'cursor-default': 64, 'aider-basic': 32, 'raw-anthropic': 16, 'inhouse-wrapper': 64}
weakest: raw-anthropic

full-screen editor opens — close anytime to keep reading.

read, then continue.